I think this is an important topic, and I hope that collectively we can generate ideas to tackle this paradox, (may be to too strong a word?).  The paradox is this: people who seek low price products often get disproportionally lower service (=output x longevity) from the payment in comparison from well made products, therefore over time may pay higher amount of money for the same service.  But this rational thinking has not stopped people from doing so, particularly for the cash constrained.
A PAYG model could solve the upfront cash issue, and allow the consumer to make chices from a wider range of products.  The question is if this can be done conveniently, cheaply and reliably.
We are all familiar with coin-operated electricity meter or laundry-mart style shared utility, but I think for a low-cost solar lighting, or even solar home system, product, convenience of operation and low collecting cost must be ranked high in the design of such a system.
I personally favor cell-phone based payment solution, where the payment-confirmation protocol is well established.  What remains is a way to reliably activate a product upon payment confirmation. I am sure each product manufacturer has its own implementation ideas, right?
Common challenges remain in many areas, delinquency definitely among them.  Ultimately, some sort of credit system is needed, and here I think a PAYG scheme would work best when managed through micro-financing community.
Of course there is also the reverse side of the problem: how would a consumer trust a payment system that is established by a product company?  Here again we could learn from, eve build on, established cell-phone banking systems in some countries like Kenya.
Rental, phased payments and financing models can all be considered variations to a general PAYG theme, I would suggest.
Just some quick thoughts...look forward to hearing more from this group!
Huashan